Table 2.
The steps in the mediation analysis investigating whether the change in pain from the pre-test to the post-test could be predicted by change in pain in the second manipulation trial, or by the second capsule effectiveness ratings.
| Dependent variable | Predictor variable | R2 | β | p |
|---|---|---|---|---|
| STEP 1 | ||||
| Placebo analgesia 45 s | Manipulation trial 45 s | 0.32 | 0.57 | 0.006 |
| Placebo analgesia 1 min 45 s | Manipulation trial 1 min 45 s | 0.34 | 0.58 | 0.005 |
| STEP 2 | ||||
| Capsule effectiveness | Manipulation trial 45 s | 0.44 | −0.66 | 0.0007 |
| Capsule effectiveness | Manipulation trial 1 min 45 s | 0.14 | −0.37 | 0.11 |
| STEP 3 | ||||
| Placebo analgesia 45 s | Manipulation trial 45 s + capsule effectiveness | Whole model: 0.35 | 0.015 | |
| Manipulation trial 45 s | 0.32 | 0.41 | 0.12 | |
| Capsule effectiveness | 0.26 | −0.23 | 0.26 | |
| Placebo analgesia 1 min 45 s | Manipulation trial 45 s + capsule effectiveness | Whole model: 0.20 | 0.11 | |
| Manipulation trial 45 s | 0.20 | 0.46 | 0.11 | |
| Capsule effectiveness | 0.08 | 0.02 | ||
Pain was recorded at 45 s and 1 min and 45 s after onset of each pain stimulus, and these recordings are analyzed separately. In the first two steps, separate regression analyses were run for each predictor. In the third analysis, both predictors were regressed on the dependent variable, as recommended by Baron and Kenny (1986).